Applying analysis technology for paddy patch detection with multiple image sources./AFA

In order to improve the outcomes of paddy rice fields through satellite images classification process. This method establishes a decision-making mechanism for uncertain data, and the improving the accuracy of the overall interpretation results. The proposed method integrates the techniques of machine learning, deep learning, and uncertainty analysis to use Ensemble Learning for information fusion to determine the final paddy rice category. Then, it introduced an automated classified process, and also simplified the process to the stage of implementation. Applying this analysis technology for paddy patch detection, the overall accuracy of rice classification can reach 92.41% on the material of optical images and 93.54% on the material of radar images. Eventually, it provides better overall results that can be obtained on ideal image quality and higher image resolution.